All good questions, but I'm afraid I have not seen a lot of good answers. We use yield history for VR N and I know there are lots of people that think that is a futile effort. I think it at least partially works in our case because we really do not have a much different soil types, it is more topsoil depth (shallow claypan soils). I think if you went from sand to muck the n mineralization might be significantly different but not the case with us.

We use a flat N credit across the field for soybeans - probably is wrong. Our problem is some years we might have 40 bpa vines and 15 bpa beans. With good early growth 15 bpa may well have as much nodulation as 40 bpa beans. There is probably some validity to varying the N credit for yield but as I see it the variability for us could easily vary more widely from year to year than from variability of yields within a field in any single year. So for me trying to "fine tune" N credits is too much like trying to throw darts at a moving dart board - I ain't too good at a stationary one so not even going to try a moving one.

N is pretty much a SWAG at best anyway - at least the way we do it - some people are probably more professional. We use long term yield vs rate studies from our immidiate area for a base rate then "tweek" them up from there for our high producing areas. So we do not have to try and manually figure out each field (been there done that takes too much energy) we have devised a set of formulas in MapCalc that gets us in the range of N rates we want to be so the process is somewhat automated - that's very important when you start doing a hundred or so fields.

Nothing at all precision about the way we do it, just hoping it is a few bucks an acre better than blanket rate.

Sorry, I can't help you much.

John